research-agent

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly lists websearch and webfetch as tools and requires in its workflow to "Gather candidate sources" and "Synthesize findings," which indicates the agent will fetch and interpret open/public (potentially user-generated/untrusted) third‑party web content that could carry instructions influencing its actions.